SE250:April 9

SOFTENG 250 Meeting 9 April 2008

Screencasts

• Screencast organisation

Dumping lots of screencasts on the same wiki page is not a good idea. Stick with one screencast per page.

Putting your screencasts on your User page is not a good idea

Screencasts need to be linked into the Work In Progress area, where people can find them

A directory of screencasts would be useful

• Filming technique

Keep the screen size as small as you can

Capture just the portion of the screen you need, usually a window or part of a window

Keep it brief

Cover just one point

Remember that the viewer can always go back and watch it again

• Don't email any further screencasts to John Hamer. Put the SWF file on your intranet site, and email the URL (e.g. http://studwww.cs.auckland.ac.nz/~UPI/UPI-n.swf) to J.Hamer@cs.auckland.ac.nz. I will copy the SWF file to the SE250 resources page, and then email you (so you can remove the file if you need to free up your file quota).

• See http://www.cs.auckland.ac.nz/tech-support/index.php/How_to_setup_your_university_intranet_webpage

• Watch the screencasts

If you come across a mistake, let the author know by putting a note on the wiki page

When there are several screencasts on the same topic, put a note next to the one you think is the best.

Labs

• I have posted comments to all the lab reports that had appeared by Tuesday evening.

• The comments are generally quite critical of your work. Many of the comments also apply to earlier labs. We are now nearly half-way through the course, and it is time to lift your game.

• The labs are designed to challenge you with some material that you are not familiar with. This is intentional. This is absolutely necessary. You are expected to face these challenges, not to sidestep them.

• Pasting program output into your report without giving any though to what the numbers might mean is sidestepping.

• You are asked to write a reflective essay. If you don't understand something, your report needs to say so. Recognising that something is important (such as the meaning of the stats tests) is an achievement in itself. You deserve (some) credit for realising this. Your report also needs to say how you intend to go about achieving an understanding. If you don't know something, you need a strategy to find out. If you don't have a strategy to find out, you need a strategy to find a strategy.

• The lab maintainers' report needs more work. It must be technically accurate, and also capture common difficulties faced by the class. The current report is not good enough.

Minutes

• Generally The lecture today consisted of the two topics above - Screen casts and The Labs.

Screen casts

• John went over all the basic things for good Screen casting, Overall the point was to keep the screen casts short and concise. Some points that are not covered above are as follows:

• Some sort of Feedback should be placed on each Screen cast, so that the author knows what things he/she can improve on.

• Mistakes should be noted on Screen casts so that they can be corrected and the correct information can be presented instead.

• Emailing the Screen casts to John takes too long and so they now should be posted up on the INTRANET using the link above.

Labs

• John went over all the basic things we need to do for the lab. Some points that are not covered above are as follows:

• Specific points about the Lab held on the 8th were made, Like specifying why the sample size was chosen.

• Also the labs should be done to a better level, they shouldn't be just about the results but also about research into the topic of the lab. The hyper text book would be a good point to start on the link below: https://www.se.auckland.ac.nz/courses/SOFTENG250/resources/HTTB-2007/SE250%20Hypertextbook.htm, but everything should be referenced.

• The lab maintainers should be working to a better level, and should be evaluating all the types of results in the lab.

Statistical analysis

All definitions are not concrete, they are just main ideas about the topic

• Entropy - is the same as Compression , Its a measure of the information and context of the data. for example Normal intuition tells you that A will be followed by B, entropy will tell you the amount of randomness in the data and the chance of getting a value in compression to the rest of the data.

• Chi- square - Counts each occurrence of data.for example if you roll a dice 00 times you would expect a 1 to come approx 1/6 of the time but not exactly 1/6th of the time. This test is to check whether something happens too often or not often enough. Things that are too good to be true.

• Monte- Carlo pi - takes two numbers and gets to probability of getting both of them in a unit square.

• Algorithmic mean - is the mean of a group of numbers.

• Serial correlation coefficient - Is used to test numbers that are successively repeated over a time period.

In General

• We are about half way into this course and so self study should be done as much as possible and the hyper text book should be implemented more into your studying.

• The test date is now posted up and will be held after the mid semester break.

• More things on the agenda should be put up and discussed.

End of meeting.

Minutes taken by : asin185